Referring now to prior art knee endoprostheses, and in particular to the prior art knee prostheses with patello-femoral replacement, it has been observed that such prior art prostheses have poorly designed patello-femoral interfaces in that they do not provide reasonable congruent patello-femoral contact or sliding engagement over any appreciable range of knee motion.
More particularly, such prior art prostheses typically produce contact stresses which result in yielding and fatigue of the plastic bearing surface typically present in such prostheses. This result is caused by the fact that the bearing surface of the femoral component, over which the patella prosthesis must pass, generally has several regions or segments of differing shape. For example, there is typically a fairly long, singly curved segment blending into a first doubly curved segment blending again into a second, and different, doubly curved segment. These varying segments or regions provide the femoral portion of the femoral-tibial articulation, and those segments or regions do not have a common generating curve. Thus, when the patella prosthesis goes through its excursion over the femoral articular flange, the patella prosthesis undergoes a variety of contact conditions, namely, substantial portions of line contact, portions of point contact, and perhaps limited portions of area or congruent area contact. As is known, line contact and point contact conditions generally produce high contact stresses which produce yielding and substantial wear of plastic prostheses. Hence, the extended wear life needed for successful prosthetic implantation is not realized.
Referring next to typical prior art tibio-femoral knee prostheses, it has been observed that those prior art knee prostheses which allow axial rotation and anterior-posterior motion in addition to flexion-extension motion have incongruent contact (usually theoretical point-contact) between the femoral and tibial bearing surfaces, producing excessive contact stresses leading to deformation and/or early wear and undesirably short prosthetic life. Also, wear products have been shown to produce undesirable tissue reactions which may contribute to loosening of the prosthetic components.
Those prior art knee prostheses which do provide congruent or area bearing contact fail to provide the needed axial rotation, or when cruciates are present the needed anterior-posterior motion. This lack of axial rotation and anterior-posterior motion has been shown clinically and experimentally to result in deformation and loosening of the tibial components, and such prostheses now appear to be falling into disuse.
Current prostheses of the dislocatable cruciate retaining type, such as the Geomedic knee replacement shown in U.S. Pat. No. 3,728,742 issued Apr. 24, 1973 to Averill et al., that produce area contact provide only one axis of rotation relative to the femur for the flexion-extension motion. Normal flexion-extension is, however, characterized by a polycentric flexion-extension motion where rotation relative to the femur occurs about many axes. This polycentric motion, which results from the action of the cruciate ligaments and condylar shape, allows for more efficient utilization of muscle forces by providing a posterior shift of the axis when effective quadriceps action is important and an anterior shift when hamstrings effectiveness is important. Furthermore, in the human knee it is this polycentric action, and the shape of the posterior condyles, which influence this motion so as to allow full flexion capability for the knee. Failure to provide appropriate knee geometry inhibits, when cruciate ligaments are present, this natural polycentric motion and thus tends to restrict muscle effectiveness and inhibit flexion. These restrictions tend to increase both loading on the prosthesis (which increases wear or likelihood of deformation or breakage) and loading between prosthesis and bone (which increases the possibility of component loosening).
Other knee designs, such as the Townly type, avoid overconstraint by introducing incongruency of the articulating surfaces. The incongruency, while necessery to avoid overconstraint, unfortunately results in instability and excessive contact stresses.
It is further believed that loosening problems result from the direct attachment of plastic prosthetic components to bone through the use of relatively brittle cement that is weak in tension. Specifically, it has been demonstrated that even relatively thick plastic components when loaded in a normal fashion produce undesirable tensile stresses in the acrylic cement commonly used to secure such plastic components to bone. Such loading tends to produce bending of the plastic component which causes the ends of the plastic component to lift away from the bone, thereby subjecting the bone-cement attachment to tension. As is known, cement has very poor tensile fatigue properties. The bone to which the plastic prosthesis is cemented also appears to be adversely affected by tensile loads. Accordingly, it is believed that these combined effects contribute substantially to prosthetic loosening problems and, specifically, it has been noted where clinical failure due to loosening occurs in a knee prosthesis that it is almost always the plastic prosthesis component which loosens.
Another prior art prosthesis problem exists with regard to knee endoprostheses for implantation in those cases wherein the cruciate ligaments are functionally absent but where the collateral ligaments are functional or at least reconstructable. In the absence of cruciate ligaments, the prosthetic replacement must provide anterior-posterior knee joint stability so as to replace that stability otherwise provided by the cruciates. Until recently most such cases were treated by a stable hinge-type knee prosthesis which, unfortunately, appears to suffer from the loosening problems described above and furthermore typically produces substantial bone loss as a result of the relatively great bone resection required for implantation. Necrosis of the bone, caused by altered mechanical bone stresses, is also a problem with the hinge-type knee prostheses. More recent attempts have been made to treat such cases with surface replacement prostheses such as the prostheses known as the Total Condylar and similar knee prostheses. However, these knee prostheses have theoretical point-contact bearing surfaces with their above-noted attendant problems and, in addition, such prostheses tend to have instability and dislocation problems which result, at least in part, from these point-contact bearing surfaces.
Where the cruciate ligaments are present, most surgeons would prefer their retention, since they provide important internal stabilizers and, together with the condylar geometry of the femur and tibia, control the rotation axis of the knee. Furthermore, these ligaments provide anterior-posterior (A-P) stability. Thus, it is desirable to reserve the cruciate ligaments, even though reasonable stability can be provided by a properly designed full platform type prosthesis.
In addition, the action of the cruciate ligaments produces a shift in the rotation axis of the knee which may result in more efficient muscle utilization. Thus, preservation of these structures may provide better physiological function after knee replacement.
Still, it is not clear that the physiological advantages gained in retaining the cruciates outweigh the disadvantages of the design compromises, such as increased bearing surface incongruency and reduced tibial prosthesis bearing area, required to retain these ligaments. Thus, the desirability of retaining the cruciate ligaments in the cases of bicompartmental and tricompartmental replacement is not well established. The design described herein, however, eliminates or compensates for these design compromises, thus allowing the benefits of cruciate retention with minimal or no apparent loss in the abiliy of the prosthesis to withstand the loads to which it is subjected.
In unicompartmental replacement, the cruciates must be retained in any event since there is insufficient stability in their absence with a unicondylar replacement. Thus, for such cases a design which accommodates the cruciate ligaments is necessary.
Unicompartmental replacement with a proper bearing design allows surgical restoration of a single diseased compartment, rather than the sacrifice of normal structures to replace all three compartments of the knee. Further, reducing the number of compartments replaced has the effect of reducing prosthesis wear products. Recent evidence strongly suggests that these wear products produce adverse physiological response to the prosthesis, including an increased tendency for the prosthesis to loosen from its boney attachment.
A recent experimental knee concept, the Oxford knee, appears to provide a partial solution to the problem of overconstraint while maintaining congruency by the use of meniscal floating elements. Unfortunately, this knee suffers from several design problems which appear to limit its usefulness. The present invention, the New Jersey Meniscal Bearing Knee Replacement (NJMBK) utilizes similar concepts in an improved fashion in order to avoid some of the anticipated difficulties of the Oxford design.
The Oxford knee is shown in FIGS. 1A and 1B. The femoral components 101 consist of two metal spherical segments, each of constant radius. Bearing inserts 102 are circular in shape with a shallow spherical superior surface and a flat inferior surface. The tibial onlays 103 consist essentially of two flat plates with fixation by means of a fin 104 at the medial edge of each such flat plate.
There are several serious problems with the design of the Oxford knee of FIGS. 1A and 1B. The most basic problem is the potential for dislocation of bearing inserts 102 resulting from the limited flexion range of the device. As can be seen from FIGS. 2A and 2B the design provides excellent congruent contact up to about 90.degree. flexion. Beyond that point a surface of constant radius cannot provide proper contact within the geometric constraints imposed by having to fit the prosthesis to the human knee. Flexion substantially beyond 90.degree. produces edge contact and resulting deformation and possible dislocation of bearing inserts 102. Although 90.degree. of flexion is satisfactory from a functional standpoint, it is impractical to limit motion to this range, since activities will be encountered (such as sitting onto a low chair, or returning to the standing position after sitting in a low chair) where flexion substantially exceeds 90.degree..
The problem of insert dislocation is made more severe by axial rotation of the knee, as is shown in FIGS. 3A and 3B. In FIG. 3A, there is shown the position of bearing inserts 102 at 90.degree. flexion, but with no axial rotation of the knee. In FIG. 3B there is shown the position of bearing inserts 102 at 90.degree. flexion, but with 15.degree. (solid lines) and 30.degree. (dashed lines) of axial rotation as well. There is a pronounced overhang of bearing inserts 102, with resultant risk of dislocation, under the combination of 90.degree. flexion and 30.degree. axial rotation of the knee.
Normal distraction of one compartment of the knee during the swing phase of walking, as depicted in FIG. 4, also leaves bearing insert 102 of the prior-art Oxford knee free to dislocate.
A further disadvantage of the Oxford knee arises from the shallowness and placement of the arcs of the contact surfaces, as can be seen from FIGS. 5A and 5B. In FIG. 5A there is shown a normal knee joint, with the anatomical ramp height designated 105. Note, in FIG. 5B, that the Oxford prosthesis ramp height 106 is substantially less than the anatomical ramp height 105, and therefore the Oxford prosthesis provides less than normal medial-lateral stability. Thus, when medial-lateral shear loads are encountered, additional stress is placed on the cruciate ligaments, which may be already compromised by bone resection. Furthermore, such loading, in conjunction with flexion or extension, will produce undesirable rubbing between the edges 107 of bearing inserts 102 and the cut edges 108 of the tibial bone.
Other weaknesses of the Oxford design include lack of accommodation for patella replacement, and tibial plateau components with relatively poor load-bearing properties, as will be described later.
An alternate embodiment of the Oxford knee which attempts to deal with the problem of dislocation is depicted in FIGS. 6A-D. Unfortunately, this design has several deficiencies which make it unworkable, at least with materials now commonly used for such components. The anterior-posterior (A-P) travel limit is greatly restricted compared to that of the present invention. There is substantial unsupported area 109 of plastic bearing insert 102, as can be seen from the cross-sectional view of FIG. 6C. Flexure of the plastic bearing insert 102 will occur, transferring load to the remaining areas and thus greatly increasing bearing compressive stresses. High stress will occur in the inner cavity at the head of retaining pin 110, particularly at the edge of retaining pin 110 and at the contact between the end of retaining pin 110 and the inner cavity, as can be seen from the cross-sectional view of FIG. 6D. Furthermore, the use of retaining pin 110 makes installation of the bearing element difficult after implantation of femoral and tibial components, since it is necessary to separate the knee joint by stretching the ligaments an amount equal to the pin height in addition to the separation normally required to install bearing inserts 102.